Re: Cool URIrefs?

Dan Brickley writes:
> * Danny Ayers <danny666@virgilio.it> [2003-08-13 14:45+0200]
> > 
> > Question : Cool URIs don't change, but what about cool URIrefs?
> > 
> > In the context of RDF, I don't think this is an issue - the URIref isn't
> > expected to return something when you point a browser at it, so there's an
> > air gap between the reference and the thing being referred to.
> > 
> > But through HTML glasses, is the expectation that an anchor will always
> > refer to the same information item? Through XML glasses, same question for
> > URIs + frag ids.
> 
> Content negotiation ensures that, in the general case, they can't. #foo
> means something different (denotes, etc) depending on the content-type
> of what you get back.

I think you're reading to much formal meaning into RFC 2396, but that
may be beside the point.

In practice (and in practical theories, I think), URIRefs should be
used consistently over time and context, like all other URIs.  Do do
otherwise diminishes their value.

     -- sandro